Comprehensive Quantitative Assessment Method for Operational Risks of Autonomous Commercial Vehicles

Abstract

This invention relates to the field of intelligent vehicle technology, specifically disclosing a comprehensive quantitative assessment method for the operational risks of autonomous driving commercial vehicles. The method includes the following steps: acquiring assessment information of autonomous driving commercial vehicles; calculating assessment scores for autonomous driving commercial vehicles based on the assessment information; obtaining the risk level corresponding to the assessment score; using the assessment information corresponding to assessment scores of the same risk level as target information; generating target vectors based on the target information of individual autonomous driving commercial vehicles; clustering the target vectors to obtain clusters; obtaining representative vectors from each cluster; generating disposal strategies based on the representative vectors; and executing the disposal strategies on all autonomous driving commercial vehicles corresponding to the target vectors in the clusters. This invention improves the accuracy and comprehensiveness of risk assessment, as well as the refinement of disposal measures after risk assessment.

Description

Technical Field

[0001] This invention relates to the field of intelligent vehicle technology, specifically to a comprehensive quantitative assessment method for the operational risks of autonomous commercial vehicles. Background Technology

[0002] Driven by both technological advancements and commercial value, autonomous vehicles are gradually transitioning from testing and demonstration to commercial applications, and the commercialization process of autonomous vehicles is accelerating.

[0003] With the continuous iteration of autonomous driving technology and the constant expansion of application scenarios, the operational risks of autonomous driving commercial vehicles are showing new characteristics such as complexity and dynamism, which puts forward higher requirements for the accuracy of risk assessment and the targeted nature of handling.

[0004] In existing technologies, most approaches rely on scores obtained from parameter quantification for appropriate processing. For example, a score corresponding to a low-risk situation might warrant routine monitoring. However, this approach suffers from several drawbacks. First, it ignores the variability in risk composition. Even with the same score, the specific causes of the risk can differ significantly. Applying a uniform approach leads to imprecise risk management and fails to address the root causes of risk. Second, it neglects transportation risks, focusing primarily on driving risks and lacking operational characteristics, making it difficult to accurately assess the risks of commercial vehicles. Third, it fails to fully utilize accident information to correct for potential risks. Summary of the Invention

[0005] The purpose of this invention is to provide a comprehensive quantitative assessment method for the operational risks of autonomous commercial vehicles, thereby solving the aforementioned technical problems.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A comprehensive quantitative assessment method for the operational risks of autonomous commercial vehicles includes the following steps:

[0008] Obtain assessment information for autonomous driving operating vehicles. The assessment information includes risk information and accident information. Risk information includes risk type and corresponding occurrence frequency. Risk type includes transportation risk and driving risk. Accident information includes the number of accidents and the accident score for each accident. The accident score is determined based on the accident level.

[0009] The first score is obtained by weighted summation of driving risks and their corresponding occurrences, and the second score is obtained by weighted summation of transportation risks and their corresponding occurrences. The first score is corrected by accident information to obtain the third score, and the second and third scores are summed to obtain the evaluation score.

[0010] Obtain the risk level corresponding to the assessment score, use the assessment information corresponding to the assessment score of the same risk level as the target information, and generate a target vector based on the target information of a single autonomous driving operating vehicle.

[0011] Clustering of target vectors yields clusters, representative vectors within a single cluster are obtained, and a handling strategy is generated based on these representative vectors. The handling strategy is then executed on all autonomous driving operating vehicles corresponding to the target vectors within the clusters.

[0012] As a further aspect of the present invention: obtaining the risk level corresponding to the assessment score includes:

[0013] Set evaluation score thresholds P1 and P2, where P1 < P2;

[0014] When P < P1, it is judged as a low-risk level;

[0015] When P1≤P<P2, it is determined to be a medium-risk level;

[0016] When P2≤P, it is judged as a high-risk level.

[0017] As a further aspect of the present invention: obtaining the target vector includes:

[0018] The responsibility ratio of the autonomous driving operation vehicle at the time of the accident is obtained, and the impact score of the accident is obtained by multiplying the responsibility ratio by the accident score.

[0019] Then the target vector XL = (A1, A2, ..., A4, B1, B2, ..., B5, C1, C2, ..., C n );

[0020] Where A1 represents the number of occurrences of the first type of transportation risk, B1 represents the number of occurrences of the first type of travel risk, and C... n This represents the impact score when the nth accident occurs.

[0021] As a further aspect of the present invention: obtaining the representative vector includes:

[0022] Within a single cluster, the cosine of the angle between any two target vectors is greater than a preset value;

[0023] Within a single cluster, the sum of the cosines of the angles between the target vector a and the other target vectors is used as the selection value for the target vector a.

[0024] The target vector corresponding to the maximum screening value is used as the representative vector.

[0025] As a further aspect of the present invention, the generation and disposal strategy includes:

[0026] Group the first four dimensions of the representative vector into one group, group the five dimensions between the 5th and 9th dimensions of the representative vector into one group, and group the remaining dimensions of the representative vector into one group.

[0027] The ratio of the value of each dimension in the representative vector to the magnitude of the representative vector is used as the target value, and the average target value of each group is obtained respectively.

[0028] Sort the average target values ​​in descending order of size, and label the groups according to the sorting order as primary risk group, secondary risk group, and supplementary risk group;

[0029] The target values ​​in each dimension of the primary risk group are sorted by size to obtain the primary risk sequence, and the secondary risk sequence and supplementary risk sequence are obtained.

[0030] As a further aspect of the present invention, the generation and disposal strategy also includes:

[0031] Obtain the operation and maintenance records of autonomous driving vehicles when risks occur in the same dimension, and use text clustering methods to extract central descriptions to build a dimension-action comparison table.

[0032] Following the order of primary risk sequence, secondary risk sequence, and supplementary risk sequence, read the corresponding actions one by one, remove duplicate content, and keep the original sequence order. Then, piece together the obtained text line by line to form a maintenance strategy draft.

[0033] Perform semantic consistency checks on the maintenance strategy draft, and then convert the maintenance strategy draft into the maintenance strategy text, which is the disposal strategy.

[0034] As a further aspect of the present invention, the generation and disposal strategy also includes:

[0035] If a target value is less than a preset target value threshold, the corresponding action will not be read.

[0036] As a further aspect of the present invention: the handling strategy for all autonomous driving operating vehicles corresponding to all target vectors in the cluster includes:

[0037] If the risk level is low, then execute the first 1 / 3 of the actions in the response strategy;

[0038] If the risk level is medium risk, then execute the first two-thirds of the actions in the response strategy;

[0039] If the risk level is high, then all actions in the response strategy will be executed.

[0040] The beneficial effects of this invention compared to the prior art are as follows:

[0041] 1) This invention enables effective identification of autonomous vehicle transportation risks and driving risks, and fully couples the risk characteristics of autonomous vehicles with those of commercial vehicles; it also enables risk identification of the operational safety level and transportation safety level of autonomous vehicles.

[0042] 2) This invention integrates transportation process risks and driving behavior risks (and their weights), and comprehensively incorporates multiple dimensions such as accident frequency, accident severity, and liability attribution. Based on this, by constructing a dynamic coupling mechanism between driving risks and accident data, it significantly enhances the accuracy and characterization capability of assessing risk exposure levels in actual operating scenarios of autonomous driving commercial vehicles.

[0043] 3) This invention performs cosine clustering on the target vectors of vehicles with the same risk level, selects representative vectors, calculates target values ​​by dimension grouping, and generates primary, secondary, and supplementary risk sequences. Then, it automatically splices maintenance drafts by combining dimension-action comparison tables to form differentiated handling strategies for the entire cluster of vehicles, which greatly improves the pertinence of risk handling and operation and maintenance efficiency. Attached Figure Description

[0044] The invention will now be further described with reference to the accompanying drawings.

[0045] Figure 1 This is a flowchart illustrating the comprehensive quantitative assessment method for the operational risks of autonomous driving commercial vehicles according to the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Please see Figure 1 As shown, this invention provides a comprehensive quantitative assessment method for the operational risks of autonomous commercial vehicles, comprising the following steps:

[0048] Step 1: Obtain assessment information for autonomous driving commercial vehicles;

[0049] The assessment information includes risk information and incident information. Specifically, regarding risk information:

[0050] 1. Transportation Risk - Taking the wrong route (including lane) or going beyond the fence: The management platform can set the operating area or route of autonomous vehicles. The on-board terminal detects the location and trajectory of autonomous vehicles through various means such as vehicle positioning and roadside monitoring. When it detects that an autonomous vehicle has taken the wrong route (including lane) or gone beyond the fence, it records the location information and sends an alarm to the management platform.

[0051] 2. Transportation Risk - Exceeding the Rated Passenger Capacity. The vehicle-mounted terminal can detect the number of people in the vehicle in real time through vision and liveness detection. It is mainly for taxis. When the vehicle is overloaded, the vehicle-mounted terminal records image and video information to provide in-vehicle alerts and sends an alarm to the management platform.

[0052] 3. Transportation Risk - Speeding. The onboard terminal can recognize speed limit signs, and the management platform can issue speed limits for road sections and areas. The management platform can also issue temporary speed limits to the onboard terminal based on events such as traffic accidents and pedestrian crossings. When the autonomous vehicle exceeds the lowest of the three speed limits, the onboard terminal records the speeding information and sends an alarm to the management platform.

[0053] 4. Transportation Risk - Cargo Spillage. The onboard terminal can identify the cargo box condition of the autonomous truck, collect images of the cargo box from multiple angles at a specific frequency, and compare them with the images at the time of loading completion. When the comparison results are inconsistent, indicating that cargo has been spilled from the autonomous truck, the terminal can record image and video information and send an alarm to the management platform.

[0054] 5. Driving Risk - Lane Change Collision Risk. After the autonomous vehicle activates its turn signal, when the wheels cross the lane line, the onboard terminal can detect the distance between the autonomous vehicle's target lane and the vehicles in front and behind. If a collision risk is determined, an alarm will be sent to the management platform.

[0055] 6. Driving Risk - Unstable Vehicle Speed ​​Control. The onboard terminal monitors data such as average deceleration, average deceleration rate of change, and average acceleration in real time. When data exceeds the threshold, an alarm is sent to the management platform.

[0056] 7. Driving Risk - Unstable Lateral Control. The onboard terminal collects the standard deviation of lateral displacement within the lane within 2 seconds. If the standard deviation exceeds the threshold, an alarm is sent to the management platform.

[0057] 8. Driving Risk - Following Too Closely. When the speed of the autonomous vehicle exceeds 30km / h, the onboard terminal monitors the following distance and TTC (Time to Charge) with the vehicle in front in real time. If the following distance exceeds the threshold, an alarm is sent to the management platform.

[0058] 9. Driving Risk - Exceeding Usage Conditions. The vehicle terminal can match the input vehicle usage conditions, including lighting and weather. When it detects that the road environment exceeds the vehicle's usage conditions, the vehicle terminal will send an alarm to the management platform.

[0059] The specific alarm parameters are shown in Table 1 below;

[0060] Table 1:

[0061]

[0062] Among them, V 自 V 后 denoted as the speeds of the vehicle behind and the current vehicle in the target lane when changing lanes, respectively; N represents the total number of samplings; and Di represents the lateral displacement at the i-th sampling.

[0063] Regarding the accident information, specifically:

[0064] Accident scores are determined based on accident levels, and the values ​​are shown in Table 2.

[0065] Table 2:

[0066]

[0067] Step 2: Calculate the assessment score based on the assessment information;

[0068] The calculation model for the assessment score is as follows: ;

[0069] Where R is the assessment score, A j w represents the number of times the j-th type of transportation risk occurs. k B represents the weight of the j-th type of transportation risk. k m represents the number of times the k-th type of driving risk occurs. k C represents the weight of the k-th driving risk. x The value of X represents the impact score of the xth accident, which is the product of the accident score of the xth accident and the liability division ratio. For example, if an autonomous driving operating vehicle bears 30% of the responsibility in an accident, then the liability division ratio is 30%. X represents the total number of accidents.

[0070] Step 3: Obtain the risk level corresponding to the assessment score;

[0071] In a preferred embodiment of the present invention, obtaining the risk level corresponding to the assessment score includes:

[0072] Set evaluation score thresholds P1 and P2, where P1 < P2;

[0073] When P < P1, it is judged as a low-risk level;

[0074] When P1≤P<P2, it is determined to be a medium-risk level;

[0075] When P2≤P, it is judged as a high-risk level.

[0076] It should be noted that the above calculation formula first multiplies the occurrence frequency of the four types of transportation risks by their respective weights and sums them to obtain the static risk at the transportation level. Then, it multiplies the occurrence frequency of the five types of driving risks by their corresponding weights and sums them. The formula is then modulated by an amplification factor that is "one plus the sum of the impact scores of all accidents". This means that when the number of accidents is higher and the liability ratio is greater, the driving risk is amplified proportionally, fully reflecting the coupling effect of accidents on the overall risk.

[0077] This weighting and amplification mechanism can compress multi-source heterogeneous risks onto a continuous score on the same scale, with a higher score representing a higher risk.

[0078] Based on this, two thresholds are set from small to large to form three intervals: low, medium, and high. On the one hand, this ensures that the score and risk level are strictly monotonous, and on the other hand, a buffer zone is left between the thresholds to reduce misjudgments caused by measurement errors or occasional fluctuations.

[0079] After this processing, the comprehensive score not only retains the differences in the composition of different risks, but can also be directly used for subsequent clustering and strategy matching, so as to achieve accurate classification and targeted handling of operating vehicles.

[0080] Step 4: Use the assessment information corresponding to the assessment scores of the same risk level as the target information, and generate a target vector based on the target information of a single autonomous driving operating vehicle;

[0081] In another preferred embodiment of the present invention, obtaining the target vector includes:

[0082] The responsibility ratio of the autonomous driving operation vehicle at the time of the accident is obtained, and the impact score of the accident is obtained by multiplying the responsibility ratio by the accident score.

[0083] Then the target vector XL = (A1, A2, ..., A4, B1, B2, ..., B5, C1, C2, ..., C n );

[0084] Where A1 represents the number of occurrences of the first type of transportation risk, B1 represents the number of occurrences of the first type of travel risk, and C... n This represents the impact score when the nth accident occurs;

[0085] Step 5: Cluster the target vector to obtain clusters, and obtain the representative vector in each cluster;

[0086] In another preferred embodiment of the present invention, obtaining the representative vector includes:

[0087] Within a single cluster, the cosine of the angle between any two target vectors is greater than a preset value;

[0088] Within a single cluster, the sum of the cosines of the angles between the target vector a and the other target vectors is used as the selection value for the target vector a.

[0089] The target vector corresponding to the maximum screening value is used as the representative vector;

[0090] Step 6: Based on the representative vector, generate a disposal strategy and execute the disposal strategy on all autonomous driving operating vehicles corresponding to the target vectors in the cluster.

[0091] In a preferred embodiment of the present invention, the generation and disposal strategy includes:

[0092] Based on the representative vectors in the clusters, the first four dimensions (A1 to A4) are assigned to the transportation risk group, the fifth to ninth dimensions (B1 to B5) are assigned to the driving risk group, and the remaining dimensions (C1 to C5) are assigned to the travel risk group. n These are categorized into the accident impact group. The grouping is based on the fact that these dimensions correspond to the transportation process, the driving process, and the accident consequences in terms of business logic, and are homogeneous risk factors.

[0093] The magnitude of the vector is calculated, and the target value is obtained by dividing the value of each dimension by the magnitude. This ratio essentially measures the contribution rate of the risk of that dimension to the overall risk, so that dimensions of different magnitudes can be compared on the same scale.

[0094] The average of the three target values ​​is taken. The average value represents the combined contribution of all risk factors in the group. The larger the average value, the more prominent the overall risk impact of the group.

[0095] Then, the three groups are sorted according to their average value and labeled as the primary risk group, secondary risk group, and supplementary risk group, thereby determining the order of risk management.

[0096] Within the primary risk group, each dimension is further arranged according to the target value to obtain the primary risk sequence. Similarly, the secondary and supplementary risk sequences are obtained. This step further refines the risk factors within the group, providing a clear priority for action invocation.

[0097] By retrieving historical operation and maintenance records and extracting the central descriptions that best represent the general public's experience from the text corresponding to each dimension, a comparison between dimensions and operation and maintenance actions can be formed. This operation, by leveraging the natural aggregation characteristics of similar texts, can summarize high-frequency and effective handling measures from a large number of records.

[0098] The actions are queried one by one in the order of primary, secondary and supplementary risk sequence. Actions that appear repeatedly in different sequences are removed to avoid redundancy. Then, the actions are spliced ​​together in the original order to form a maintenance strategy draft. After the draft is formed, a semantic consistency check is performed. The readability and executability of the text are ensured by checking whether the context is coherent and whether the instructions are contradictory. The draft that passes the check is output as the formal maintenance strategy.

[0099] Understandably, the core purpose of this design is to map complex dimensions to a clear priority system using a grouping-normalization-sorting approach, so that the dimensions and actions that contribute the most to risk are extracted and presented in the strategy first, thereby enabling rapid targeting of high-risk factors; the operation and maintenance actions are derived from massive amounts of real records and are clustered and denoised to ensure that the measures are operable and have industry consensus; deduplication and consistency checks avoid lengthy or contradictory strategies, improving execution efficiency and readability;

[0100] It should be noted that if a target value is less than the preset target value threshold, the action in the corresponding dimension will not be read.

[0101] It should be noted that the handling strategy for all autonomous driving operating vehicles corresponding to all target vectors in the cluster includes:

[0102] Within a cluster, the risk level label of each vehicle is read, and the maintenance strategy text shared by vehicles in the same cluster is split into a sequence of actions ordered by priority. For example, the sequence includes sensor self-test, route planning refresh, software patch update, hardware connector tightening, redundant power supply detection, driving strategy retraining, etc.

[0103] Extract subsequences of different lengths according to the vehicle's risk level:

[0104] When a vehicle is marked as low risk, only the first third of actions are taken, such as performing sensor self-checks, route planning updates, and software patch updates. This allows the most common and easily accumulated hidden dangers to be addressed with minimal operational resources.

[0105] If the vehicle is classified as medium risk, the process is extended further by taking the first two-thirds of the actions and including the tightening of hardware connectors and the detection of redundant power supplies, to further cover weak links that may lead to an escalation of the risk.

[0106] When a vehicle is at high risk, all actions are executed in full, and time-consuming but crucial in-depth measures such as retraining the driving strategy are also implemented simultaneously. Through this layered and progressive invocation method, it is ensured that the actions with higher priority are triggered earlier, while the measures with lower priority but still necessary are only deployed in high-risk situations.

[0107] By using risk-driven segmented operation and maintenance, limited resources are concentrated on the most needed areas, achieving a balance between cost and safety. Low-risk vehicles do not have to bear the downtime and economic burden caused by excessive maintenance, medium-risk vehicles can promptly block potential upgrade channels, and high-risk vehicles can receive comprehensive intervention to quickly reduce the probability of accidents. Ultimately, this helps operators maintain dynamic control over the rhythm of operation and maintenance and the level of risk as the fleet size continues to expand.

[0108] It should be noted that the process of implementing handling strategies for all autonomous driving operating vehicles corresponding to all target vectors in a cluster also includes:

[0109] After calculating the comprehensive evaluation score, the weighted summation of transportation risk and the weighted summation of driving risk are extracted separately. Each of these is then divided by the final evaluation score to obtain its contribution weight, with the one with the larger weight being the current dominant risk type. If the result shows that transportation risk is the dominant risk type, the transportation safety system of the relevant enterprise is first reviewed in a targeted manner. For example, the process of reviewing carrier qualifications, dynamic loading monitoring, and shift scheduling is improved. At the same time, model features related to cargo status recognition, loading and unloading condition matching, and abnormal trip alarms are added or strengthened on the vehicle algorithm side to reduce the probability of failure in the transportation process through the synergy of the system and the algorithm. If driving risk is the dominant risk type, the focus is on reviewing the vehicle's driving scenario processing logic in the perception, decision-making, and control chain. Data playback and model retraining are conducted for dimensions such as road obstacle recognition, lane keeping, speed curve planning, and emergency braking strategies, and road test verification is carried out to ensure that the risk factors in the driving process are directly weakened.

[0110] By dynamically comparing the contribution of the two types of risks, resources can be focused on the most prominent links, avoiding inefficient investment caused by spreading efforts evenly. Transportation risks are attributed to systems and operational processes, while driving risks are more derived from algorithms and control strategies. The improvement paths for the two are completely different. Classifying and handling them can enable the solution to accurately target the root cause, thereby quickly suppressing the sources of high accident incidence and laying the foundation for the continuous improvement of the overall operational safety level.

[0111] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A comprehensive quantitative assessment method for the operational risks of autonomous commercial vehicles, characterized in that, Includes the following steps: Obtain assessment information for autonomous driving operating vehicles. The assessment information includes risk information and accident information. Risk information includes risk type and corresponding occurrence frequency. Risk type includes transportation risk and driving risk. Accident information includes the number of accidents and the accident score for each accident. The accident score is determined based on the accident level. The first score is obtained by weighted summation of driving risks and their corresponding occurrences, and the second score is obtained by weighted summation of transportation risks and their corresponding occurrences. The first score is corrected by accident information to obtain the third score, and the second and third scores are summed to obtain the evaluation score. Obtain the risk level corresponding to the assessment score, use the assessment information corresponding to the assessment score of the same risk level as the target information, and generate a target vector based on the target information of a single autonomous driving operating vehicle. Clustering of target vectors yields clusters, representative vectors are obtained from a single cluster, and a handling strategy is generated based on the representative vectors. The handling strategy is then executed on all autonomous driving operating vehicles corresponding to all target vectors in the clusters. The generation and disposal strategies include: Group the first four dimensions of the representative vector into one group, group the five dimensions between the 5th and 9th dimensions of the representative vector into one group, and group the remaining dimensions of the representative vector into one group. The ratio of the value of each dimension in the representative vector to the magnitude of the representative vector is used as the target value, and the average target value of each group is obtained respectively. Sort the average target values ​​in descending order of size, and label the groups according to the sorting order as primary risk group, secondary risk group, and supplementary risk group; The target values ​​in each dimension of the primary risk group are sorted by size to obtain the primary risk sequence, and the secondary risk sequence and supplementary risk sequence are obtained. The generation and disposal strategies also include: Obtain the operation and maintenance records of autonomous driving vehicles when risks occur in the same dimension, and use text clustering methods to extract central descriptions to build a dimension-action comparison table. Following the order of primary risk sequence, secondary risk sequence, and supplementary risk sequence, read the corresponding actions one by one, remove duplicate content, and keep the original sequence order. Then, piece together the obtained text line by line to form a maintenance strategy draft. Perform semantic consistency checks on the maintenance strategy draft, and then convert the maintenance strategy draft into the maintenance strategy text, which is the disposal strategy.

2. The comprehensive quantitative assessment method for operational risks of autonomous driving commercial vehicles according to claim 1, characterized in that, The risk levels corresponding to the assessment scores include: Set evaluation score thresholds P1 and P2, where P1 < P2; When P < P1, it is judged as a low-risk level; When P1≤P<P2, it is determined to be a medium-risk level; When P2≤P, it is judged as a high-risk level.

3. The comprehensive quantitative assessment method for operational risks of autonomous driving commercial vehicles according to claim 1, characterized in that, Obtaining the target vector includes: The responsibility ratio of the autonomous driving operation vehicle at the time of the accident is obtained, and the impact score of the accident is obtained by multiplying the responsibility ratio by the accident score. Then the target vector XL = (A1, A2, ..., A4, B1, B2, ..., B5, C1, C2, ..., C n ); Where A1 represents the number of occurrences of the first type of transportation risk, B1 represents the number of occurrences of the first type of travel risk, and C... n This represents the impact score when the nth accident occurs.

4. The comprehensive quantitative assessment method for operational risks of autonomous driving commercial vehicles according to claim 1, characterized in that, Obtaining the representative vector includes: Within a single cluster, the cosine of the angle between any two target vectors is greater than a preset value; Within a single cluster, the sum of the cosines of the angles between the target vector a and the other target vectors is used as the selection value for the target vector a. The target vector corresponding to the maximum screening value is used as the representative vector.

5. The comprehensive quantitative assessment method for operational risks of autonomous driving commercial vehicles according to claim 1, characterized in that, The generation and disposal strategies also include: If a target value is less than a preset target value threshold, the corresponding action will not be read.

6. The comprehensive quantitative assessment method for operational risks of autonomous driving commercial vehicles according to claim 5, characterized in that, The handling strategies for all autonomous driving operating vehicles corresponding to all target vectors in the cluster include: If the risk level is low, then execute the first 1 / 3 of the actions in the response strategy; If the risk level is medium risk, then execute the first two-thirds of the actions in the response strategy; If the risk level is high, then all actions in the response strategy will be executed.

Images